Timing and mechanism of conceptus demise in a complement regulatory membrane protein deficient mouse

Problem Crry is a widely expressed type 1 transmembrane complement regulatory protein in rodents which protects self‐tissue by downregulating C3 activation. Crry −/− concepti produced by Crry +/− × Crry +/− matings are attacked by maternal complement system leading to loss before day 10. The membrane attack complex is not the mediator of this death. We hypothesized that the ability of C3b to engage the alternative pathway's feedback loop relatively unchecked on placental membranes induces the lesion yielding the demise of the Crry−/− mouse. Method of Study We investigated the basis of Crry −/− conceptus demise by depleting maternal complement with cobra venom factor and blocking antibodies. We monitored their effects primarily by genotyping and histologic analyses. Results We narrowed the critical period of the complement effect from 6.5 to 8.5 days post‐coitus (dpc), which is immediately after the conceptus is exposed to maternal blood. Deposition by 5.5 dpc of maternal C3b on the placental vasculature lacking Crry−/− yielded loss of the conceptus by 8.5 dpc. Fusion of the allantois to the chorion during placental assembly did not occur, fetal vessels originating in the allantois did not infiltrate the chorioallantoic placenta, the chorionic plate failed to develop, and the labyrinthine component of the placenta did not mature. Conclusion Our data are most consistent with the deposition of C3b being responsible for the failure of the allantois to fuse to the chorion leading to subsequent conceptus demise.


| INTRODUC TI ON
The alternative pathway (AP) of the complement system is vital to host defense against pathogens yet also contributes to autoimmune and inflammatory diseases. In host defense, the effects of complement activation are largely dependent on the cleavage of C3 by the C3 convertases, enzyme complexes assembled during complement activation ( Figure 1A). C3 cleavage products opsonize targets for clearance by phagocytic cells, promote inflammation, and perturb cell surfaces via the membrane attack complex (MAC; C5b-9).
Regulation of convertase activity is required to maintain homeostasis and protect self-tissue. Membrane cofactor protein (MCP; CD46) and decay accelerating factor (DAF; CD55) are proteins expressed on most healthy cells that regulate convertases. Factor H (FH) is an abundant plasma protein that prevents formation of and dissociates convertases via decay accelerating activity and cofactor activity ( Figure 1B). Absence of these normal regulators of the AP are the cause of a number of human diseases, including atypical hemolytic uremic syndrome (aHUS), 1-6 C3 glomerulopathies (C3GN), [1][2][3][4][5][6] age-related macular degeneration (AMD), [7][8][9][10][11] and protein-losing enteropathy. 12 Crry is a widely expressed type 1 transmembrane protein in rodents that downregulates the AP. It has strong cofactor activity for C3b and moderate decay accelerating activity for the classical pathway. [13][14][15] Crry's broad expression profile and regulatory activities are similar to membrane cofactor protein (MCP/CD46) in primates ( Figure 1B). 16 MCP has a limited expression profile in rodents, being present primarily on the inner acrosomal membrane of spermatozoa. 9,[16][17][18] Crry −/− concepti produced by Crry +/− × Crry +/− matings are attacked by the maternal AP leading to loss before 10 days. 15 Notably, Crry −/− pups can be rescued if the mother is deficient in any 1 of the 4 components (C3, FB, FD, and properdin) of the AP (13)(14)(15) and X. Wu and J.P. Atkinson, unpublished). Fetal loss that occurs is not dependent on antibody (μMT −/− background) or the classical or lectin pathway of complement (C4 −/− background). 13 Moreover, demise is not mediated by the MAC, as the C6-deficient mouse conceptus does not rescue the demise phenotype. 19 However, C5a may play a minor role: C5 −/− background led to ~5% of offspring being Crry −/− although much less than the expected 25% of offspring. 13 These results indicate that fetal loss occurs primarily through APdirected events, prior to formation of the C5 convertase and the membrane attack complex.
Herein, we test the hypothesis that the ability of C3b to engage the AP feedback loop in the absence of proper membrane regulation induces the placental lesion yielding the demise of the Crry −/− mouse. Specifically, we surmise that a reduction in maternal AP activity at ~6.5 days post-coitus (dpc) would prevent conceptus demise. The mouse blastocyst enters the uterus at 4.5 dpc yet the conceptus is not directly exposed to maternal blood until 5.0-6.5 dpc. 20 As early as 6.5 but completely by 7.5 dpc, the ectoplacental cone cells that will evolve into the chorioallantoic placenta are bathed in maternal blood, thereby exposing to proteins of the complement cascade. Neutrophil depletion and C3a receptor blockade failed to rescue the conceptus. Overall, the results further suggest that C3b deposition is responsible for the loss of the conceptus. [13][14][15]19 2 | MATERIAL S AND ME THODS

| Mouse breeding and genotyping
Mice were bred and maintained under pathogen-free conditions at Washington University School of Medicine (WUSM) in St. Louis, MO in accordance with institutional animal care guidelines. The Crry knockout mouse was originally generated by Molina and colleagues 14 and has been maintained at WUSM. The Crry −/− allele was genotyped by PCR as described. 15 The C3aR knockout mouse was a gift from Richard Wetsel (University of Texas, Houston) and genotyped by PCR.

| Timed matings and harvesting embryos
After female mice were placed in the male's cage, each subsequent day the female was checked for a vaginal plug. The day of plug observance was assigned 0.5 dpc. Mice were expected to deliver at 19.5 dpc. Pregnant mice were sacrificed by CO 2 asphyxiation in accordance with institutional guidelines. The uterus was then removed and each implantation site was separated surgically. The muscular uterus was removed under a dissecting microscope while the implantation site was placed in cold PBS. Pregnant mice were F I G U R E 1 Alternative pathway of complement activation: feedback loop and regulation by cofactor activity. A, Four plasma proteins, C3, factor B (FB), factor D (FD), and properdin (P) assemble into the AP C3 convertase. C3 convertase has a continuous activity that generates a basal level of C3b. C3b engages the zymogen Factor B (FB) that is then cleaved by the serine protease Factor D (FD) to generate Bb and Ba. Bb remains bound to C3b while Ba is liberated (not shown). C3bBb is an active, albeit transient, AP enzyme complex that cleaves C3 to form C3a and C3b. C3a is a small peptide anaphylatoxin when released. C3b covalently binds to a nearby target (T) surface forming an ester linkage. Properdin binds to and stabilizes the C3bBb complex, increasing the half-life 5-to 10fold. Thus, the cleavage of C3 to C3b may result in a potent positive feedback amplification loop. The AP C5 convertase (C3b) 2 BbP (not shown) cleaves C5 to C5a and C5b. C5a is a potent anaphylatoxin. C5b binds C6 and C7 and C5b,6,7 complex attaches to a membrane. The binding of C8 followed by multiple C9s then completes the formation of the membrane attack complex (MAC), which has the capacity to perturb cellular membranes including generating pores to lyse cells. B, Crry mediates membrane cofactor activity in the mouse. After Crry binds to C3b, then Factor I (FI), a plasma serine protease, can now cleave C3b to iC3b. iC3b has no hemolytic potential because it does not engage FB to initiate the feedback loop. Crry has intrinsic complement regulatory activity in that it primarily acts on the same cell on which it is expressed 32 routinely sacrificed at day 11.5 dpc. Implantation sites, including the concepti and extraembryonic membrane surrounded by decidua, were weighed to confirm the dpc. Genotyping was performed on each litter. To accomplish this, the conceptus was removed, washed 7× in cold PBS in a microtiter plate, and digested in proteinase K (20 μg/mL) overnight at 55°C. DNA was precipitated, suspended in 10 mmol/L Tris, 0.1 mmol/L EDTA, and then analyzed by PCR as described previously. 15

| Immunohistology
Implantation sites were fixed in 10% formalin overnight, embedded in paraffin, 8 μm sections were rehydrated, antigen retrieval was performed, non-specific binding was blocked, and specific immunostaining was conducted as described below. In the case of Crry staining, antigen retrieval was performed in 10 mmol/L citric acid (anhydrous; Sigma-Aldrich), 0.05% Tween-10, pH 6.0 in a pres- For trophoblast staining, 7.5 dpc implantation sites were collected as described above. Antigen retrieval was accomplished with 10 mmol/L Tris-EDTA pH 9.0 for 3 minutes in a pressure cooker. Staining was accomplished with TROMA-I (1:50 dilution of hybridoma supernatant; Developmental Studies Hybridoma Bank, University of Iowa) and goat anti-rat light chain horseradish peroxidase (HRP). Staining was visualized with DAB. Controls employed a second Ab only.

| FACS analysis of cells derived from 7.5 dpc implantation sites
Implantation sites were harvested as described above. Each site was cut into 12 pieces and placed in RPMI 5% fetal bovine serum (FBS).

| Cobra venom factor (CVF) treatment
Cobra venom factor (Quidel, A600) was administered intraperitoneally (20 μg/mouse) with a 31G insulin syringe (Terumo). Depletion of C3 occurs in <1 hour. C3 hemolytic and antigenic activity is undetectable for up to 3-4 days and then there is a gradual increase to normal levels over approximately a week. 21,22

| Neutrophil depletion
Neutrophils were depleted by intraperitoneal injection (IP) injection of RB6-8C5, a rat IgG2b mAb against Gr-1 (Ly6G/C). A 250 μg dose of this Ab depleted neutrophils in the periphery for 5 days and a 500 μg dose depleted for 6 days. Both doses are followed by a rebound neutrophilia (approximately a doubling of pre-depletion levels), which we were unable to overcome with an additional IP dose 4 days after the first dose. 1A8 is a second anti-Gr-1 rat IgG2a Ab (BioXCell). A 500 μg dose of 1A8 depleted ~50% of the neutrophils when the peripheral blood was assayed at 72 hours.
In the initial experiments, we used RB6-8C5 that was a gift from Emil Unanue (Washington Univ. School of Medicine, Department of Pathology and Immunology, St. Louis, MO). 23 RB6-8C5 was also produced within the laboratory by hybridoma cells. The mAb was purified from supernatants on a protein G column and then dialyzed against PBS. RB6 was also purchased from BioXCell (West Lebanon, NH).

| Timing of AP activity required for conceptus loss and depletion of the AP by CVF
Crry −/− products of conception are known to undergo demise before 10.5 dpc with maternal AP components playing a critical role. [13][14][15]24 We sought to determine when the lethal AP activity occurred and we chose treatments that could suspend AP activity after implantation. CVF is a C3b analog that forms a stable AP convertase with Bb (CVFBb), and unlike host AP C3 convertases, CVFBb is not susceptible to inhibitory activity of the complement regulators. 25 Consequently, treatment with CVF rapidly depletes circulating C3 and diminishes FB, fully exhausting complement activity. 22 We depleted AP by injecting 20 μg CVF IP into newly pregnant mice and examined Western Blots of serum for C3. This approach yielded no detectable C3 for the next 4 days, with C3 levels returning to 50% of normal by 7 days post-injection. The fetuses were genotyped and the placenta examined at 11.5 dpc. CVF treatment administered between 3.5 and 5.5 dpc rescued Crry −/− mice, compared to controls (Table 1). Treatment at 6.5 or 7.5 dpc led to ~50% rescue. Of note, CVF treatment had no adverse effect on pup survival, as Crry −/− pups were born to CVF-treated mothers similar to controls (Table 2), matured as expected, and Crry −/− females were fertile (not shown).

| Histology of the implantation site
The above results showed that conception loss was prevented if AP inhibition began before 6.5 dpc, and a partial effect existed if inhibition began by 6.5-7.5 dpc. The ectoplacental cone cells of the mouse are precursors to the labyrinthine trophoblast in the placenta. The cone cells contact maternal RBCs as early as TA B L E 1 Treatment with CVF prior to 8.5 dpc rescues Crry −/− implantations a dpc of treatment (litters)

Resorbed implantations (% of total)
Crry −/− implantations rescued 3 and resorbed in the absence of CVF treatment. Implantation sites were evaluated and genotyped at 11.5 dpc. **P < .01 and *P < .05 compared with the proportion of full size implantations if not treated. The 5.5 dpc had only 1 treated litter; a statistical test was not performed. Partial rescue indicates the number of resorbed implantations was between a full rescue and no rescue. TA B L E 2 Treatment with CVF enables the birth of Crry −/− pups a 5.5 dpc and are bathed in blood at 6.5 dpc. 20 The yolk sac placenta offers nutrition to the conceptus until 9.5 dpc, when the labyrinthine placenta assumes the major role in maternal-fetal exchange ( Figure 2). Importantly, we observed that at 8.5 dpc there was a substantial difference in the size of the allantoic vessels, a lack of proliferation of the labyrinthine trophoblasts, and smaller than control embryos in Crry −/− gestations ( Figure 3). Moreover, the Crry −/− implantations at 9.5 dpc failed to evolve like controls, with an undeveloped labyrinth and unexpanded chorioallantoic vessels ( Figure 4).

| Neutrophils are not required for Crry-/conceptus loss
In the APLS model of embryonic lethality, 30 complement-mediated damage is dependent on neutrophils. We observed neutrophils surrounding Crry −/− implantation sites on 7.5 dpc, but the impact on conception viability has not been established. 13 We pursued this the sites of implantation at 7.5 dpc (Figure 6). Examination of the surviving concepti at 10.5 and 13.5 dpc (Table 5), along with genotyping of litters (Table 6), indicated that neutrophil depletion by RB6-8C5 did not improve Crry −/− viability. Similarly, depletion of neutrophils by treatment of pregnant mice with an anti-Ly6G-specific Ab (1A8, 500 μg/mouse) had no detectable effect (Table 6). We concluded that neutrophils are not essential for loss of the Crry −/− embryos.

| D ISCUSS I ON
The AP of the complement system is a constant sentinel, activating continuously on surfaces and in the fluid phase. 15,31 Regulators on cell membranes and in plasma are essential to control the level of activation. Crry is widely expressed on the surface of mouse cells and carries cofactor activity. 32 In plasma and on cellular debris, factor H (FH) performs a similar role. Crry is able via CA to permanently stop AP activation on cell membranes. 33 In the Crry −/− mouse model, excessive AP activation leads to conceptus demise by 8.5 dpc of development, highlighting the critical role played by regulators of the AP. Crry −/− pups can be rescued by down modulating the levels of the AP activating proteins in the mother. 15 The data herein show that transient depletion of the AP, either with CVF or an anti-properdin Ab, is sufficient to rescue Crry −/− concepti. There is a critical window at 6.5-7.5 dpc when AP activation leads to conceptus loss. In the case of CVF, treatment prior to this window produces full rescue, but treatment during this 24-hour period rescues about one-half of the concepti. Interestingly, if CVF is given at 3.5 dpc prior to implantation of the blastocyst into the decidua, maternal C3 levels rise to ~50% by 7.5 dpc (4 d after CVF treatment) but Crry −/− embryos survive. This replicates the finding F I G U R E 6 Neutrophils are present around 7.5 dpc embryos and RB6-8C5 depletes them from around embryos. A, Neutrophils are present around the embryo at 7.5 dpc (anti-Gr-1 staining; 200×). B, This staining is specific for Gr-1 + and is not present in the isotype control. C, At higher magnification (400×), the nuclear pattern characteristic of neutrophils can be seen inside of Gr-1 + cells. D,E, Identical patterns are observed in staining for Gr-1 + (RB6-8C5, D) and Ly6G (1A8, E). F, Ly6G staining is specific to 1A8 and not seen in the isotype control. G-I, Neutrophil depletion with 500 μg RB6-8C5 at 4.5 dpc leads to absence of neutrophils in the tissue at 7.5 dpc of Gr-1 + (G) and Ly6G + cells (H,I). em, embryo; epc, ectoplacental cone that full blockade of the AP is not required for embryo survival, as haploinsufficiency of AP activating components will also rescue the concepti. 15

| Timing of C3 deposition
C3 deposition on the ectoplacental cone of the Crry −/− mouse occurs at approximately 7.5 dpc. 13,14 We propose this takes place because there is no membrane-based complement inhibitor on these cells.
Decay-accelerating factor (DAF), a GPI-anchored regulator, is not expressed in the labyrinthine placenta until after ~10.5 dpc. 13 Other C3 regulators of complement such as FH are present in maternal blood but they are unable to compensate for Crry deficiency. For example, FH is normal in Crry −/− mouse but this complement regulator is insufficient to limit the diffuse AP activation observed on cells in the Crry −/− mouse. 15 We surmise that the underlying cause of conceptus loss in this model is a dysregulated AP of complement activation on cell membranes.